The Consumption, Income, and Wealth of the Poorest

[Pages:68]Public Disclosure Authorized

Public Disclosure Authorized

Policy Research Working Paper

WPS7337 7337

The Consumption, Income, and Wealth of the Poorest

Cross-Sectional Facts of Rural and Urban Sub-Saharan Africa for Macroeconomists

Leandro De Magalhaes Raul Santaeulalia-Llopis

Public Disclosure Authorized

Public Disclosure Authorized

Development Research Group Poverty and Inequality Team June 2015

Policy Research Working Paper 7337

Abstract

This paper provides new empirical insights on the joint distribution of consumption, income, and wealth in three of the poorest countries in the world -- Malawi, Tanzania, and Uganda -- all located in Sub-Saharan Africa (SSA). The first finding is that while income inequality is similar to that of the United States (US), wealth inequality is barely one-third that of the US. Similarly, while the top of the income distribution (1 and 10 percent) earns a similar share of total income in SSA as in the US, the share of total wealth accumulated by the income-rich in SSA is one-fifth of its US counterpart. The main contributions of the paper are to document: (i) this dwarfed transmission from income

to wealth, which suggests that SSA households face a larger inability to save and accumulate wealth compared with US households; and (ii) a lower transmission from income to consumption inequality, which suggests the presence of powerful institutions that favor consumption insurance to the detriment of saving. These features are more relevant for rural areas, which represent roughly four-fifths of the total population. The paper identifies the few successful pockets of the SSA population that are able to accumulate wealth by exploring sources of inequality such as age, education, migration, borrowing ability, and societal systems.

This paper is a product of the Poverty and Inequality Team, Development Research Group. It is part of a larger effort by the World Bank to provide open access to its research and make a contribution to development policy discussions around the world. Policy Research Working Papers are also posted on the Web at . The authors may be contacted at rauls@wustl.edu and Leandro.DeMagalhaes@bristol.ac.uk.

The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent.

Produced by the Research Support Team

The Consumption, Income, and Wealth of the Poorest:

Cross-Sectional Facts of Rural and Urban Sub-Saharan Africa for Macroeconomists

Leandro De Magalh~aes University of Bristol

Rau?l Santaeul`alia-Llopis Washington University in St. Louis

Universitat de Val`encia

JEL: E20; E21; O10; O55; I32 Keywords: Inequality, Consumption, Income, and Wealth, Sub-Saharan Africa

We thank Kathleen Beegle, Taryn Dinkelman, Pascaline Dupas, Francisco Ferreira, John Kaboski, Talip Kilic, and John Temple for comments and suggestions at different stages of this project. We also thank the seminar participants at the World Bank LSMS group and the SED in Seoul 2013.

1 Introduction

Cross-sectional facts on the distribution of consumption, income, and wealth (CIW) are readily available for a large set of modern industrialized economies, see the special issue of the Review of Economic Dynamics on "Cross Sectional Facts for Macroeconomists" (Krueger et al. (2010)) and the more recent studies of D?iaz-Gim?enez et al. (2011) and Piketty (2014). Such distributional facts have been extensively used to build and test macroeconomic theories that incorporate heterogeneous household saving behavior for the study of, almost invariably, rich economies (Heathcote et al. (2009)). For these macroeconomic frameworks to be useful in poor countries they need to be fully contextualized and informed by the behavior of households in these countries. Hence, a good understanding of household-level CIW inequality for poor countries is required and to date has been missing. The main contribution of this paper is to help close this gap by providing a careful and comprehensive dissection of CIW behavior in three of the poorest countries in the world in 2010 -- Malawi, Tanzania and Uganda -- using new and unique nationally representative data. To construct the whole distribution of CIW a series of technical issues had to be addressed: converting consumption from own production into kilograms, deseasonalizing consumption, and assigning a monetary value to the agricultural production that is not sold on the market.

To gain a perspective on the relative poverty of the countries investigated note that in 2010, income per capita was US$359 in Malawi, US$524 in Tanzania, and US$471 in Uganda. That is, our households live with country averages below (or close to) US$1 per capita per day, and less so in rural areas, where the overwhelming majority of the population lives: 84% in Malawi, 71% in Tanzania, and 85% in Uganda.1 For comparison purposes, in Mexico, the poorest country studied in Krueger et al. (2010), the income per capita in 2010 was US$8,920 and the population living in rural areas was 22%. In Thailand, a country extensively investigated in the development literature, these figures were US$4,802 and 56% respectively.

Our main finding is a large and widespread inability to accumulate wealth in both rural and, to a lesser extent, urban areas of Sub-Saharan Africa (SSA). First, while income dispersion in SSA is of similar magnitude to that of the US and Europe, wealth inequality in SSA is close to one-third that of the US.2 Further, while the transmission from income to wealth inequality is not apparent from the separate inspection of their respective distributions, our study of the joint behavior of income and wealth suggests a substantially lower transmission in SSA than in the

1Indeed, Malawi is currently the poorest country in the world in terms of income per capita according to the 2014 World Development Indicators (World Bank). Malawi has been disputing this unfortunate position over the past 10 years with the Democratic Republic of Congo and Niger, also sub-Saharan African countries.

2For these cross-sectional comparisons between SSA and the US we largely draw from the US statistics reported in D?iaz-Gim?enez et al. (2011) and Piketty (2014).

2

US. This phenomenon is clearly starker in the rural areas than in the urban areas of SSA: we find that the top 1% income-richest households in rural and urban SSA, respectively, hold 4% of total wealth in rural areas and 11% in urban areas, while this figure is 26% in the United States.3 That is, the income-richest households in rural SSA accumulate only as much as 15% of the share of total wealth that the income-richest US households are able to accumulate in terms of US wealth. The equivalent figure for urban areas in SSA is 46%, which also suggests a reduced ability to accumulate wealth in SSA than in the US. The behavior of the joint density of income and wealth strengthens this message as we find significant positive correlations between income and wealth only for the top 20% of the income distribution. The correlation is stronger in urban areas, 0.30, than in rural areas, 0.12, while this figure is much larger, 0.57, for the entire US economy.

The direct exploration of household saving provides further insights. We find that only households in the top 1% and 5% of the income distribution in rural and urban areas are able to save, and they do so with high saving rates similar to those of the top income earners in the US. One way to relate this finding (i.e., saving rates similar to those in the US for the top income earners in SSA) with the result discussed in the previous paragraph (i.e., lower share of total wealth held by the top income earners in SSA than in the US) is through the lesser persistence of high income in SSA than in the US. Indeed, we find that within a span of four years, 51% of rural households at the top quintile of the income distribution fall to a lower quintile, as do 46% of urban households, while this figure is only 23% in the US.4 That is, households in the top quintile of the income distribution in SSA leave that quintile twice as fast as their counterparts in the US, and this downward mobility occurs somewhat faster in rural than urban areas. This lesser persistence of high income -- and hence, of saving -- helps to explain the low transmission from income to wealth and the low wealth accumulation in SSA compared with that in the US, even for the top income earners who actually save.

Interestingly, despite the generalized inability to save that we find in SSA, we find a concurrent lower transmission from income to consumption inequality in SSA than in the United States. Precisely, economy-wide, the inequality of consumption is 47% that of income (in terms of the variance in logs). The figure for the United States is larger, roughly 55% (Blundell et al. (2008)). Furthermore, the joint density of consumption and income implies a correlation between consumption and income of 0.30 in rural areas and 0.53 in urban areas, suggesting that this insurability of consumption is larger in rural areas than in urban areas. Combining our two findings -- lower transmission from income inequality to both wealth inequality and consumption

3For brevity, the precise numbers reported in the introduction refer to Malawi. 4We use data from Uganda for these mobility statistics, as it has a panel component not available for Malawi.

3

inequality in SSA than in the US -- suggests that the greater ability to contain the dispersion in consumption in SSA compared with the US must occur through insurance mechanisms other than self-insurance (i.e., own savings).

The higher ability to insure in rural areas is further revealed from direct evidence on selfreported information about shocks, coping strategies, and ability to borrow. Among rural households 71% report being hit by a shock in the past 12 months (most commonly weather-related shocks), and among urban households 39% report a shock (most commonly high food prices). Despite the higher occurrence of shocks in rural areas, of the rural households that report a shock, 51% have some form of insurance to cope with the shock, while this figure is smaller in urban areas, 40%. The ability to insure consumption can also depend on the ability to borrow. In this case, conditional on needing a loan, urban households show higher application rates, 40%, than rural households, 27%.5 While this urban-rural differential in the ability to borrow suggests urban households might be better able to self-insure, we note that in urban areas borrowing is used 3.6 times more often for start-up capital than for consumption, while this ratio is 1.6 for rural areas. This suggests that such borrowing is used mostly for production and not necessarily for consumption insurance.

In decomposing CIW we find unambiguous signs of a subsistence economy where food consumption requirements can severely limit the ability to save. Specifically, the share of food in total consumption is above 50% throughout the income distribution, except for the top 1% in rural areas and the top 5% in urban areas. Coincidentally, these two groups are the only ones with positive saving rates across the income distribution. The composition of income (mostly agricultural) in rural areas also points towards a subsistence economy. Only in the top 1% of the income distribution in rural areas does the share of agricultural income fall below 50%. The similarity in the wealth portfolio across income groups (except for the top 1%) further reflects an inability to accumulate wealth. The only components of wealth that show signs of accumulation are livestock and non-housing durables.

While the overall power to accumulate wealth is dismal in the SSA countries we consider, the examination of cross-sectional sources of inequality allows us to identify pockets of wealth accumulation. First, over the life cycle, wealth in the rural areas of SSA (i.e., mostly land) follows a hump pattern that increases, in Malawi, by a factor of 2.65, while this increase is by a factor of 7.13 in urban areas, further suggesting a greater ability to accumulate wealth in urban areas than in rural areas. However, again, even the urban increase is small compared to the life cycle behavior of wealth in the United States that increases by a factor of roughly 20. Second, we find

5As we discuss below, these figures deal directly with self-selection as households are asked whether they needed a loan and whether they applied or not, independently of the need.

4

that while education tends to allow for more income and consumption, it does not necessarily imply higher wealth accumulation, particularly in rural areas. In contrast, in urban areas, the more-educated households (i.e., "secondary education or more") are overrepresented in the top 1% of the wealth distribution. Indeed, only the more-educated urban households have positive savings on average. Third, CIW averages in rural areas and urban areas are very similar across countries. The large discrepancy is between CIW averages between rural and urban areas. Indeed, the difference in average CIW between rural and urban areas is an order of magnitude larger than the difference in average CIW across countries. Such large differences should trigger important rural-to-urban migration flows, which we do observe as 60% of current urban household heads report having migrated from rural areas. We find that while rural-to-urban migration improves their consumption and income, we also find that rural-to-urban migrants are not better able to accumulate wealth than the local households in the hosting urban region. The only migration group with a positive saving rate is the one formed by urban-to-urban migrants, who represent 18% of the urban population. Finally, we find that the ability to accumulate wealth can also be related to gender and to the particular societal systems a household belongs to (i.e. matrilineal or patrilineal). In particular, for Malawi, the only group with average positive savings is the the one formed by matrilineal households in urban areas in the center of the country.

Our paper relates to a vast literature in development economics. First, our results on the inability to save and accumulate wealth in Malawi, Tanzania, and Uganda are directly related to the experimental results in Dupas and Robinson (2013a,b) for Kenya on savings constraints.6 Similar to Rosenzweig and Wolpin (1993) findings for India, we find that the households in rural areas of East Africa that save do so in livestock and other nonhousing durables. This may be due to a lack of alternative saving options. Second, in terms of consumption insurance, Townsend (1994) and Udry (1994) suggest the presence, respectively, of informal arrangements in villages in India and northern Nigeria. These results confirm our findings of a weaker transmission from consumption to income inequality in SSA than in the United States, and in rural rather than urban areas. The phenomenon that the ability to insure consumption is stronger in rural areas than in urban areas has been previously explored in Morten (2013) and Munshi and Rosenzweig (2014) for India, and in Santaeul`alia-Llopis and Zheng (2015) for China. Two aspects set our work apart from previous studies. The first is the use of nationally representative data, which is a natural approach from our macroeconomic perspective. The second, and most important, is that we can establish the joint behavior of the triplet of CIW and document -- as far as we know for the first time -- the joint phenomenon of a low ability to accumulate wealth (i.e., generate growth in the economy) and a low transmission from income to consumption (i.e., a high ability

6See also the recent review in Karlan et al. (2014).

5

to insure consumption).

The rest of the paper is sectioned as follows. In Section 2, we provide a full account of the data construction. In Section 3, we discuss in detail the distributional facts of CIW for Malawi, Tanzania, and Uganda, with more emphasis on Malawi. The emphasis on Malawi is due to the sample size being four times larger and the data more detailed than for the other two countries. The corresponding tables for Tanzania and Uganda are available in the online appendix. In Section 4, we decompose the consumption basket, the income sources, and the wealth portfolio in rural and urban areas. In Section 5, we investigate the sources of inequality and identify pockets of the population with positive savings. In Section 6, we explore economic mobility using panel information for Uganda in 2005 and 2009. We conclude in Section 7.

2 ISA Data and Measurement Issues

The Integrated Surveys on Agriculture (ISA) are conducted under the umbrella of the Living Standards Measurement Study (LSMS). ISAs greatly improve previous LSMS household surveys in use for decades. In particular, ISAs are unique in the level of detail in CIW for households that make their living through agriculture. The ISAs allow us to recover -- for any practical purpose --the entire deseasonalized household-level budget constraint in a set of the poorest countries of SSA.7 However, in addition to the ISA improvements in data collection and availability, a correct measurement of CIW requires further adjustments. In particular, due to the importance of agricultural production, estimating the monetary value of income and consumption requires us to convert physical units in which production and consumption are reported into a single unit, assign prices to items that are produced by the household for own consumption, and deseasonalize consumption. As we show, the choice of which price to use to value non-sold production -- either the price at the gate (i.e., production price) or consumption price -- can have large impacts on the estimated monetary values of agricultural production.

2.1 Units Conversion of In-Kind Items: From Pails to Kilograms

In household surveys from poor countries, it is standard to report amounts of consumption (including from own production), production (mainly agricultural) and additional sources of income (e.g., gifts and transfers) in units that are not necessarily harmonized across time or space. For example, in the Malawi ISA, households are asked to report the amount they produce of a given item in any unit they wish, and this varies from bags, dishes, bunches, and pails, to kilograms

7To the best of our knowledge, the availability of the CIW triplet is rather unique as these three items are not supplied at the same time by any other single dataset, not even in developed countries. See also the discussion in Heathcote et al. (2009).

6

 ................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download